67.
38 The Central Angle of the Regular 17-Gon
Author(s): James J. Callagy
Source: The Mathematical Gazette, Vol. 67, No. 442 (Dec., 1983), pp. 290-292
Published by: Mathematical Association
Stable URL: http://www.jstor.org/stable/3617271
Accessed: 03-11-2015 02:53 UTC
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact [email protected].
Mathematical Association is collaborating with JSTOR to digitize, preserve and extend access to The Mathematical Gazette.
http://www.jstor.org
This content downloaded from 137.149.200.5 on Tue, 03 Nov 2015 02:53:14 UTC
All use subject to JSTOR Terms and Conditions
290 THE MATHEMATICAL GAZETTE
(7) If the top row consists of an arithmetic sequence al, a2, ... with
common differenced, then
Sk
and the term (x + k)c in f(x) is replaced by
(dx + a)(x + k)c-'.
Thus for:
2 - 8 11 -4-1 17 202- 26 29 32 35-8- 41 ...
,, 1024- 38,8- 84 113 145 -4-8 221 ...
- -48 132 245 9 611 ...
132 7-- 988 ...
-4-- 1120 .
u,= 2 so f,(x)=(3x+2)(x+ 1),
ua= 5 so f2(x)=(3x + 5)(x + 2)(x + 1),
ua- 8 so f(x) (3x + 8)(x + 3)(x + 2),
u4 = 13 so f4(x) - (3x + 11)(x + 4)(x + 3)(x + 1)2.
JOHN G. SLATER
Beechen Cliff School, Bath
67.38 The central angle of the regular 17-gon
The possibility of inscribing a regular n-gon in a circle using straight edge
and compasses only was a problem of fundamentalimportance for Greek
mathematicianswho succeeded in solving it only for values of n of the forms
2k+2; 2k.3; 2k 5; 2k 15.
The general solution and real mathematical significance of the problem,
however, were not to be revealed until late in the eighteenth century when
Gauss proved that the construction was possible, if and only if the odd
prime factors of n were of the form 22k + 1, that is, Fermat primes. He
showed its feasibility for k = 2, by solving an equation of the seventeenth
degree, without bothering to carry out the actual construction of the
resulting 17-gon, and wrote down the expression:
27r
cos - 1= /17 - 1 + V(34 - 2/17)1
17
'/[17 + 3V/17- V/(34 - 2V/17) - 2(34
+ + 2v/17)1.
This content downloaded from 137.149.200.5 on Tue, 03 Nov 2015 02:53:14 UTC
All use subject to JSTOR Terms and Conditions
NOTES 291
It was the first entry in a diary he began on 30th March 1796 just one
month before his nineteenth birthday and his pride in the discovery is
commemoratedby an engraving of the figure on a monumenterected at his
birthplacein Brunswick.It was widely acclaimed as a historic breakthrough
in a subject which had been dormant for over two thousand years but the
more interestingaspect of this discovery is perhaps, the traditionthat it had
motivated Gauss to devote his life to mathematics.
The actual construction of the regular 17-gon has attracted little or no
attention in modern times but many classical constructions were given by
leading mathematicians in the nineteenth century like J. Serret, K. Von
Staudt, F. Richelot, O. Hermes while that of W. H. Richmond in 1893 is
still being presented in modern texts as probably the simplest of them; see,
for example, Coxeter's Introductionto geometry (1969). Having appliedthe
Lemoine criteriaand standardfor 'simplicity'of a Euclidean construction,I
am presenting the following construction as not only simpler but also as a
more direct geometrical representation of the expression attributed to
Gauss. I consider that the direct construction of the central angle is more
relevant than just finding the third and fifth vertices of the polygon itself as
Richmond did.
R
Q - _ ^, - . P
Not very rational
'High arithmetricprecision is often called for in this kind of product: divide 22 by 7 and pi
is approximatedto 10 significant digits, more than enough for most people." From a review
of a computer in Practical computing (sent in by Robert Fairthorne).
This content downloaded from 137.149.200.5 on Tue, 03 Nov 2015 02:53:14 UTC
All use subject to JSTOR Terms and Conditions
292 THE MATHEMATICALGAZETTE
Given a unit circle with rectangularaxes PQ, RS the central angle 2z/17
requiresthe construction of ten points as shown:
(i) A on OR such that AO = 4.
(ii) B, C on OP such that AB, AC are the internal and external
bisectors of angle OAP.
(iii) D, E on OP such that CD = CA and BE = BA.
(iv) M, the mid-pointof QD.
(v) F on OS such that MF = MQ.
(vi) G on semi-circlediameterOE such that OG = OF.
(vii) H on OP such that EH - EG.
(viii) P1 on the original unit circle where P, H is perpendicularto OP.
Then angle POP, is 27r/17.
A proof that this construction does produce the requiredangle involves
simple calculations with radicals using standard Euclidean theorems, but it
is not easy. On the other hand, it may be verified with a calculator that
0.9324722 is a good approximationfor both cos POP1 and cos 27r/17.
JAMES J. CALLAGY
UniversityCollege, Galwai'
67.39 The location of twin primes
In this note we derive a test for locating twin pairs of primes 6N- 1 and
6N + 1, N being a positive integer,a test that is both easy to apply and most
suitable for use with a computer. The test does not involve determining
separatelythe prime nature of each of 6N - 1 and 6N + 1, and can be used
either for one particular value of N, or for all values of N up to a stated
maximumvalue.
We use the terminology that if 6N- 1 and 6N + 1 are both primethen N
is twin prime. For example 2 is twin prime,but 4 is not. It is well known that
there are very large twin primes, such as 12 015, and much larger,of course.
An unsolved problem of number theory is whether there are indefinitely
large twin primes, and though the present analysis gives some idea of the
decreasing likelihood of N being twin prime as N increases, we do not claim
to have added anything significantto the major conjecture concerning twin
primes.
Our starting point is the observation that the quadratic Q(N, M; x) given
by
Q(N,M; x) = (6N + 1)x2 + 12Mx + (6N -- 1)
factorises over the integers for M = N and M = 3N2, but if 6N - and
6N + 1 are both prime it fails to factorise over the integers for any integer
value of M such that N < M < 3N2; and conversely, if one or other of
6N- I and 6N + 1 is composite then there exists at least one integer value
This content downloaded from 137.149.200.5 on Tue, 03 Nov 2015 02:53:14 UTC
All use subject to JSTOR Terms and Conditions
